Advances in Planning and Scheduling of Remote Sensing Instruments for Fleets of Earth Orbiting Satellites

This paper describes a system for planning and scheduling science observations for fleets of Earth observing satellites. Input requests for imaging time on an Earth observing satellite are specified in terms of the type of data desired, the location to be observed, and an objective priority of satisfying the request. The problem is to find a sequence of start times for observations and supporting activities such as instrument slewing and enforcement of instrument thermal duty cycles, that satisfy a set of temporal and resource constraints describing the physical operation of the spacecraft. We assume that there are more requests that can possibly be serviced over a given scheduling window, and that images may vary in their scientific utility, leading to an optmization problem. This paper presents an approach to solve this problem employing 1) a formal declarative model of the problem, 2) stochastic sampling methods to find plans, and 3) special purpose heuristics based on a generalized contention measure.